Closed danrgll closed 1 month ago
Can you give an illustration or example or snippet of how you are visualizing the final feature?
Exemplary idea:
run_args:
# run_pipeline provided via neps.run()
# or
run_pipeline:
path: "<path_to_run_pipeline>"
name: "<name_of_run_pipeline>"
pipeline_space: "pipeline_space.yaml"
# or
pipeline_space:
epochs:
lower: 1
upper: 6
log: True
optimizer:
choices: [ "adam", "sgd", "adamw" ]
constant:
value: 3
root_directory: "my_directory"
max_evaluations_total: 20
post_run_summary: True
searcher: "bayesian_optimization.yaml"
# or
searcher:
algorithm: bayesian_optimization
initial_design_size: 7
surrogate_model: gp
acquisition: EI
log_prior_weighted: false
acquisition_sampler: random
random_interleave_prob: 0.1
disable_priors: false
prior_confidence: high
sample_default_first: false
Closes #96
To streamline configuration management, we could consolidate multiple YAML configurations—such as pipeline_space, custom_searcher, and others—into a single run_args file. Instead of referencing these configurations, they would be defined directly under their respective keys within run_args. This approach would simplify setup and can enhance usability by allowing users to manage all settings from one centralized file.
Note: This consolidation could be optional, not removing the existing functionalities.